Closed ghost closed 8 years ago
Reminder to add things to etherpad -- @whyrusleeping @lgierth @mappum @richardlitt @dignifiedquire @davidar @chriscool @rht @diasdavid
hey @diasdavid @whyrusleeping -- reminder: your things o/ o/
sorry, just added :)
This week i had a bunch of small things. Lots of CR upkeep, figuring out various issues, lost of discussions with people.
Etherpad for the next sprint, https://github.com/ipfs/pm/issues/36: https://etherpad-mozilla.org/sprint-36
Some very small fixes in sharness tests:
And improved shellquote():
Apologies for missing today's checkin, I thought I would be able to do it in Newark Airport, but the airport was impressively full today and after waiting in all the queues, had to run to my gate.
My main focus for this past week was working on a interopable go-libp2p and get ipsurge off the ground.
Usage: ipsurge COMMAND [OPTIONS]
Available commands:
browse Open your application in a browser
init Initialize a ipsurge project
preview Preview your application through a collection of snapshots
publish Publish your project
versions Check each version published
@davidar update: https://github.com/ipfs/apps/issues
I chatted with @whyrusleeping about logrus and logging for ipget, and decided against it: the logging isn't very human friendly, and CLI tools like this are primarily talking to humans. go-logging looks nice, but I think just writing to stderr is sufficient for this tool.
I didn't get to delve very deeply into testing, but I got a wealth of information about how current ipfs get
testing works with sharness, which I'll tackle this week.
@noffle If you git blame
the code, the one who put logrus into almost all logging was me, sorry.
One reason against logrus that I know of is that logrus doesn't tell line number when an err happens (a much requested feature as in https://github.com/Sirupsen/logrus/issues/63).
Though it would be more helpful if you elaborate on "the logging isn't very human friendly", lest the back-and-forth switching between the two libs.
(will fix the composite files PR later)
I chatted with @whyrusleeping about logrus and logging for ipget, and decided against it: the logging isn't very human friendly, and CLI tools like this are primarily talking to humans. go-logging looks nice, but I think just writing to stderr is sufficient for this tool.
Curious what you mean here? You may want to write a custom formatter if it's to show users and not the developers.
@Sirupsen its not really that its not 'human friendly' its just information overload to use a full fledged logging library for an app that has functionality similar to 'wget'.
While youre here... any news on the filename/line-number front? logrus is definitely more flexible, but the lack of line numbers makes debugging more difficult for me (prompting the discussion about switching back to go-logging
)
Curious what you mean here? You may want to write a custom formatter if it's to show users and not the developers.
Jeromy did a better job of clarifying than I did -- sorry about that. I didn't know about the line number feature-want, but yes, my qualm was with logrus being heavy (creating a custom formatter that outputs essentially just what fmt.Fprintf(os.Stderr, "...")
does) in relation to the tininess of this tool. Having line numbers (maybe behind a -d
debug switch) would be nice, though.
@whyrusleeping I've never needed the functionality myself and noone has ever opened a convincing PR to do it that works in every situation (see this comment). I'm very surprised noone has written a hook for it, it doesn't necessarily have to be in core.
Let's move this to an appropriate issue in another repo
Sprint Goals
Sprint Discussions
See https://github.com/ipfs/pm#sprint-discussion-schedule
Please take notes in a separate pad, if you can, and link it here.
Sprint Deliverables
@yourhandle
@noffle
@jbenet
@amstocker
python-ipfs-api:
@mappum
@RichardLitt
@lgierth
@diasdavid
@dignifiedquire
[ ] IPFS
(Ref https://github.com/ipfs/logo/issues/6)@rht
ipfs object cache $hash
/ipfs object get --cache-only $hash
? i.e. store a particular hash into blocks without outputting to $PWD or stdout. Useful for partial backup of parts of a huge node, pre-ipfs-cluster@whyrusleeping